The increase of anthropomorphisationanthropomorphisation in technological applicationstechnological applications, particularly in chatbotschatbots based on generative AIgenerative AI, has led to a growing interest in their discursive characteristicsdiscursive characteristics and limitslimits. Anthropomorphism, understood as the attribution of human traits to non-human entities, can manifest itself in both the form and behaviourbehaviour of chatbots, and language occupies a central role in this process. While recent advances have enabled chatbots to mimic human language with unprecedented linguisticlinguistic accuracylinguistic accuracy, questions remain about the sense of authenticityauthenticity of these interactions, as well as the ethicalethical and emotional implications they entail. This study aims to identify the specific characteristicscharacteristics of anthropomorphic discourseanthropomorphic discourse generated by AI chatbots and to establish the limits of their imitative ability. Through the analysis of 16 conversationalconversational interactions with two of the leading large language model-based tools, CopilotCopilot and GeminiGemini, the imitation of human language was evaluated at three levels: superficial, situational, and quality and relevancerelevance. The findings allow us to advance the definition of discursive anthropomorphismanthropomorphism in these systems and discuss its implications for the future of human–machine interactionhuman-machine interaction. Finally, the chapter examines the ethical and socialsocial implications of anthropomorphisation in generative AI, exploring its discursive performance and emphasising the need to reflect on the boundariesboundaries between humans and machines in the context of growing technological integration.